tasteR

This workshop is designed to give you a taste of what R can do. We’ll run through the process of importing, tidying, transforming, visualising and reporting data using the suite of packages that make up the tidyverse. These packages help lower the barrier to entry for new users and will hopefully inspire you to code regularly in R.

Learning outcomes

By the end of the workshop you’ll:

  • Be familiar with the RStudio interface
  • Read data into R and inspect it
  • Query and transform data
  • Understand how to build graphics using ggplot2

R

What is R?

R is an open source programming language for statistical analysis and data visualisation. It was developed by Ross Ihaka and Robert Gentleman of the University of Auckland and released in 1995. There are now over 14,000 packages available for R which provide functions for machine learning, genomics, time series forecasting, and interactive graphics amongst many others.

R is widely used in academia and rapidly replacing SPSS and other proprietary statistical software in undergraduate programmes. It is also used by well known companies like Google, Netflix and Airbnb for data analytics. Many graphics published by news outlets like the Financial Times and the BBC are generated in R. The UK Government is also embracing R to help make their statistical reporting workflow more efficient and reproducible.

Why use R?

  • free
  • open source
  • advanced statistical analysis
  • publication ready graphics
  • increasingly used to support undergraduate teaching of statistics
  • programming skills are attractive to employers

Since R is a language it is also:

  • open
  • shareable
  • reproducibile
  • human readable
  • diffable

and therefore supports the principles of open science.

There are also several R packages useful for those working in local government. For example, the fingertipsR package enables you to download population and public health indicators.

RStudio

RStudio is an integrated development environment (IDE) for R. It’s intuitive interface makes working with R much easier. It supports syntax highlighting, tab completion and is integrated with R Markdown.

RStudio is freely available under the GNU Affero General Public License v3. A commercial desktop license is also available.

The panes

RStudio has four different panes:


RStudio's panes
RStudio’s panes


  • The Console pane (top left) is used to execute R commands immediately.

  • The Environment pane (top right) shows the datasets, models, and plots that are loaded in the current R session. This pane also contains tabs with a scrollable history of executed code, connections to databases and Git options.

  • The Files pane (bottom right) shows plots and interactive web content, help documentation, previous commands, and R packages that you can install and load.

  • The Source pane (bottom left) appears when you open a new file e.g. File -> New File -> R Script. Code can be saved in dedicated .R scripts and executed in the console with Ctl-Enter/Cmd-Enter. Syntax highlighting and tab completion are also available.

Customisation

The appearance of RStudio can be changed to suit you:

  • Increase the font size: Tools > Global Options > Appearance > Editor Font size
  • Move the Console pane to the right: View > Panes > Console on Right
  • Add a margin column line: Tools > Global Options > Code > Display > Show margin
  • Automatic text wrapping: Tools > Global Options > Code > Editing > Soft-wrap R source files

You can also change RStudio’s overall theme. Opting for a dark theme reduces the amount of glare that your eyes are subject to. Change the global theme to Dark by selecting ‘Appearance’ in the Global Options menu and opt for an Editor theme with a dark palette such as ‘Material’.


RStudio's dark mode
RStudio’s dark mode


Keyboard shortcuts

RStudio has many useful shortcuts that enable you to keep your hands on the keyboard thereby boosting your coding productivity. For example,

  • Ctrl/Cmd + Enter: Run your selected code in the Console
  • Ctrl/Cmd + Shift + M: Add piping operator
  • Ctrl/Cmd + L: Clear the console window
  • Ctrl/Cmd + Shift + R: Add a section break
  • Shift + Ctrl/Cmd + 1: Make the console full screen
  • Ctrl/Cmd + Shift + A: Format your code

NB The use of Ctrl or Cmd depends on whether you are using a Windows or a Mac device.

For a complete list of all available shortcuts just type: Alt + Shift + K

Setup

The data

The dataset that we will be exploring derives from speedofanimals.com. These are measurements for speed, mass and length for a variety of animals living in the water, air and on land.


speedofanimals.com
Source: speedofanimals.com


The dataset contains 5 variables:

  • name: the common name of the animal
  • habitat: whether the animal lives in water, air, or on land
  • length_cm: the length of the animal in centimetres
  • mass_kg: the mass of the animal in kilograms
  • speed_kph: the top speed of the animal in kilometres per hour

You can download the dataset from here.

Organise your project

Adopting a consistent folder structure for your data analyses will help to ensure that your projects are reproducible. A project can be organised using a simple file structure like this:

project/
|
├── data/          # store your datasets
|
├── script.R       # your R script
|
└── output/        # all your plots, models etc

Ideally, raw and intermediate datasets would be separated to ensure that you don’t accidentally overwrite your data.

project/
|
├── data/
│   ├── raw/       # read-only pre-processed datasets
│   └── processed/ # intermediate datasets
|
├── script.R       # your R script
|
└── output/        # all your plots, models etc

Set your working directory

Point your R session to your project folder using: Session > Set Working Directory > Choose Directory

NB It’s not good practice to set your working directory at the top of your R script because absolute paths don’t promote reproducibility.

Optional: Set up a project in RStudio

Open a new R script

File > New File > R Script

Install R packages

You only need to install an R package once. Subsequent package updates can be handled by selecting Packages > Update in the Files pane of RStudio.

install.packages("tidyverse")

Load R packages

Packages need to be loaded at the start of every R session to give you access to the functions you need.

library(tidyverse)

Import

R can handle a range of data formats: .csv, .xlsx, .sav, .dta, .sas, .xml, .json, .geojson etc. Some data formats require specific packages.

Read a CSV

We can load a CSV file using the read_csv function with an argument which provides the file path to the dataset. A new object called speedofanimals is created which stores the data that we have loaded in the R session.

speedofanimals <- read_csv("data/speed_of_animals.csv")

Inspect the data

The glimpse() function prints all the variables, their data types and the first few values.

glimpse(speedofanimals)
## Observations: 64
## Variables: 5
## $ name      <chr> "African Bush Elephant", "African Wild Dog", "Black Ma…
## $ habitat   <chr> "Land", "Land", "Land", "Land", "Land", "Land", "Land"…
## $ length_cm <dbl> 690, 100, 340, 250, 140, 90, 50, 28, 140, 120, 3, 520,…
## $ mass_kg   <dbl> 8000.00, 30.00, 1.40, 620.00, 65.00, 18.00, 6.00, 0.13…
## $ speed_kph <dbl> 40.0, 72.5, 32.2, 35.0, 120.0, 69.0, 48.0, 20.0, 72.4,…

Coerce data types

  • character variables can store strings of characters
  • factor variables can store a fixed and known set of possible values
  • integer variables can only store whole numbers
  • double variables can store decimal numbers
  • logical variables can store TRUE or FALSE
  • date variables can store a date - time
speedofanimals <- read_csv("data/speed_of_animals.csv", 
               col_types = cols(
                 name = col_factor(NULL),
                 habitat = col_factor(NULL),
                 length_cm = col_integer(),
                 mass_kg = col_double(),
                 speed_kph = col_double()
                 ))
speedofanimals

Tidy

Tidy data are structured for use in R and satisfy three rules:

  1. Each variable must have its own column
  2. Each observation must have its own row
  3. Each value must have its own cell
speedofanimals.com
Source: R for Data Science


gather()

Convert variable names into values and structure the dataframe into a more machine readable format (wide to long)

speedofanimals_long <- gather(speedofanimals, measure, value, -name, -habitat)
speedofanimals_long

spread()

Convert dataframe into a more tabular, human readable format (long to wide)

spread(speedofanimals_long, measure, value)

Note that pivot_long() and pivot_wide() will shortly be available as more intutively named functions.

separate()

Split out values stored in one variable into multiple variables

split <- separate(speedofanimals_long, measure, c("measure", "unit"), sep = "_")
split

unite()

Join values stored in multiple variables into one

unite(split, measure, c(measure, unit), sep = "_")

Transform

select()

Pick and rename columns in a dataframe.

by name

select(speedofanimals, name, speed_kph)

by exclusion

select(speedofanimals, -speed_kph)

with a range

select(speedofanimals, 1:3)

begins with

select(speedofanimals, starts_with("hab"))

rename

select(speedofanimals, animal = name)

filter()

Subset rows in a dataframe by an expression.

less than

filter(speedofanimals, length_cm < 100)

equal to

filter(speedofanimals, name == "Roadrunner")

not equal to

filter(speedofanimals, habitat != "Air")

with regex

filter(speedofanimals, grepl("African", name))

by multiple conditions

filter(speedofanimals, mass_kg > 1000 & habitat == "Land")

by a vector

filter(speedofanimals, name %in% c("Coyote", "Roadrunner"))

arrange()

Change the order of rows in a dataframe.

ascending order

arrange(speedofanimals, speed_kph)

descending order

arrange(speedofanimals, desc(speed_kph))

mutate()

Create new variables with functions of existing columns in a dataframe.

calculation

mutate(speedofanimals, speed_mph = speed_kph * 0.62137)

new data type

mutate(speedofanimals, name = factor(name))

by multiple conditions

mutate(speedofanimals, speed_mph = speed_kph * 0.62137,
       name = factor(name))

%>%

The piping operator allows you to combine multiple operations together in a chain. For example,

speedofanimals %>% 
  filter(habitat == "Air") %>% 
  mutate(speed_mph = speed_kph * 0.62137) %>%
  select(name, speed_mph) %>% 
  arrange(desc(speed_mph))

The pipe obviates the need to create objects for each intermediate transformation of your dataframe and makes your code more human readable. For example, the code above can be translated as:

Filter rows in the speedofanimals dataframe whose values in the habitat column are equal to “air”, then create a new column which calculates the speed in miles per hour of each animal, then select the name and speed_mph columns, and then sort the resulting rows by the values of speed_mph in descending order.

The piping operator makes it easy to apply functions to dataframes grouped by specific variables. For example, the code below use the group_by() and summarise() functions to calculate the mean speed of animals by habitat

speedofanimals %>% 
  mutate(speed_mph = speed_kph * 0.62137) %>% 
  group_by(habitat) %>% 
  summarise(mean_speed_mph = round(mean(speed_mph), 1)) %>% 
  arrange(desc(mean_speed_mph))

Visualise

“A statistical graphic is a mapping from data to aesthethic attributes (colour, shape, size) of geometric objects (points, lines, bars).”
Hadley Wickham (2016)

Variable Geometry Aesthetic
mass_kg point x-position
speed_kph point y-position
length_cm point size
habitat point fill

Building a plot

Initialise a ggplot() object

ggplot()

Load the data and map variables to the x and y axes

ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph))

Add points

ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph)) +
  geom_point()

Optional: change the colour of the points

ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph)) +
  geom_point(colour = "tomato")

Map the habitat variable to fill and add a border to the points

ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph, fill = habitat)) +
  geom_point(shape = 21, colour = "black", alpha = 0.6)

Log transform the scale

library(scales)
ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph, fill = habitat)) +
  geom_point(shape = 21, colour = "black", alpha = 0.6) +
  scale_x_continuous(trans = log_trans(), breaks=c(1,10,100,10000)) +
  scale_y_continuous(trans = log_trans(), breaks=c(1,10,100))

Add a title, label the axes, include a caption, and change the legend title

ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph, fill = habitat)) +
  geom_point(shape = 21, colour = "black", alpha = 0.6) +
  scale_x_continuous(trans = log_trans(), breaks=c(1,10,100,10000)) +
  scale_y_continuous(trans = log_trans(), breaks=c(1,10,100)) +
  labs(title = "Relationship between body mass and speed \namongst selected animals",
       caption = "Source: speedofanimals.com",
       x = "Body mass (kg, log)",
       y = "Top speed (kph, log)",
       fill = "Habitat")

Change the overall theme

ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph, fill = habitat)) +
  geom_point(shape = 21, colour = "black", alpha = 0.6) +
  scale_x_continuous(trans = log_trans(), breaks=c(1,10,100,10000)) +
  scale_y_continuous(trans = log_trans(), breaks=c(1,10,100)) +
  labs(title = "Relationship between body mass and speed \namongst selected animals",
       caption = "Source: speedofanimals.com",
       x = "Body mass (kg, log)",
       y = "Top speed (kph, log)",
       fill = "Habitat") +
  theme_minimal()

Map length_cm to size and add some further tweaks

ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph, fill = habitat, size = length_cm)) +
  geom_point(shape = 21, colour = "black", alpha = 0.6) +
  scale_fill_brewer(palette = "Set2") +
  scale_size_continuous(range = c(1, 10)) +
  scale_x_continuous(trans = log_trans(), breaks=c(1,10,100,10000)) +
  scale_y_continuous(trans = log_trans(), breaks=c(1,10,100)) +
  labs(title = "Relationship between body mass and speed \namongst selected animals",
       caption = "Source: speedofanimals.com",
       x = "Body mass (kg, log)",
       y = "Top speed (kph, log)",
       fill = "Habitat",
       size = "Length (cm)") +
  theme_minimal() +
  theme(
    plot.margin = unit(rep(30, 4), "pt"),
    panel.grid.minor = element_blank(),
    axis.title = element_text(size = 8, face = "plain", hjust = 1),
    plot.caption = element_text(size = 8, hjust = 1, margin = margin(t = 15)))

Save the plot as a png at a high resolution

ggsave("scatterplot.png", scale=1, dpi=300)

Add interactivity

library(ggiraph)
speedofanimals$name <- str_replace_all(speedofanimals$name, "'", "")
p <- ggplot(data = speedofanimals,
       aes(x = mass_kg, y = speed_kph, fill = habitat, size = length_cm)) +
  geom_point_interactive(aes(tooltip = name), shape = 21, colour = "black", alpha = 0.6) +
  scale_fill_brewer(palette = "Set2") +
  scale_size_continuous(range = c(1, 10)) +
  scale_x_continuous(trans = log_trans(), breaks=c(1,10,100,10000)) +
  scale_y_continuous(trans = log_trans(), breaks=c(1,10,100)) +
  labs(title = "Relationship between body mass and speed \namongst selected animals",
       caption = "Source: speedofanimals.com",
       x = "Body mass (kg, log)",
       y = "Top speed (kph, log)",
       fill = "Habitat",
       size = "Length (cm)") +
  theme_minimal() +
  theme(
    plot.margin = unit(rep(30, 4), "pt"),
    panel.grid.minor = element_blank(),
    axis.title = element_text(size = 8, face = "plain", hjust = 1),
    plot.caption = element_text(size = 8, hjust = 1, margin = margin(t = 15)))
girafe(code = print(p))

Model

Calculate body lengths per second: (metres per second / length)

speedofanimals <- speedofanimals %>% 
  mutate(body_lengths_s_km = ((speed_kph*1000) / 3600) / (length_cm / 100))
select(speedofanimals, name, body_lengths_s_km) %>% 
  arrange(desc(body_lengths_s_km))

Calculate what speed ‘feels like’ when 6ft (1.83m) tall

speedofanimals <- speedofanimals %>% 
  mutate(rel_speed_kph = ((body_lengths_s_km * 1.83) * 3600) / 1000)
select(speedofanimals, name, body_lengths_s_km, rel_speed_kph) %>% 
   arrange(desc(rel_speed_kph))

Communicate

The rmarkdown package allows you to write dynamic reports in R. These are plain text files (.Rmd) written in markdown containing chunks of embedded R code.

Why use it?

  • Documents with embedded code are reproducible
  • Integrates the entire data analysis workflow from importing to reporting.
  • Documents can be outputted to a variety of formats: HTML, PDF, Microsoft Word, slideshows, dashboards, websites, and books.

Markdown

Markdown is a simple syntax that allows you to add formatting to plain text. By enclosing text with asterisks you can add *emphasis* or write inline `code` with backticks.

Emphasis

Use asterisks and tilde to add emphasis to your text.

**bold** → bold

*italics* → italics

***bold italics*** → bold italics

~~strikethrough~~ → strikethrough

Paragraphs

End a line with two or more spaces to create a line break. You can also use HTML break <br /> tags.

Headings

To create a header begin a line with a hashtag. Each additional hashtag makes the header smaller.

# 1st level heading   
## 2nd level heading   
### 3rd level heading   
#### 4th level heading   
##### 5th level heading

Images

Place (optional) explanatory text in brackets and the image URL or path in parentheses, preceded by an exclamation mark.

![Image: wikipedia.org](img/polar_bear.jpg)
Image: wikipedia.org

Image: wikipedia.org

To resize your image just use HTML:

<figure>
<img src="img/polar_bear.jpg" alt="A polar bear" width="30%">
<figcaption>Image: wikipedia.org</figcaption>
</figure>
A polar bear
Image: wikipedia.org

Code

Inline code
Wrap code in single backticks.

Today’s date is `format(Sys.time(), '%d %B, %Y')`

Today’s date is 08 May, 2019

Code blocks
Place 3 backticks on a line above and below the code block.

```r
ggplot(data = speedofanimals, aes(x = mass_kg, y = speed_kph)) +
  geom_point()
```

ggplot(data = speedofanimals, aes(x = mass_kg, y = speed_kph)) +
  geom_point()

Blockquotes

Blockquotes start the first line with a greater than symbol.

> There are no emergency meetings, no headlines, no breaking news. No one is acting as if we were in a crisis. Even most climate scientists or green politicians keep on flying around the world, eating meat and dairy. […] Today we use 100 million barrels of oil every single day. There are no politics to change that. There are no rules to keep that oil in the ground. So we can't save the world by playing by the rules. Because the rules have to be changed. Everything needs to change. And it has to start today. 
> (Greta Thunberg)

There are no emergency meetings, no headlines, no breaking news. No one is acting as if we were in a crisis. Even most climate scientists or green politicians keep on flying around the world, eating meat and dairy. […] Today we use 100 million barrels of oil every single day. There are no politics to change that. There are no rules to keep that oil in the ground. So we can’t save the world by playing by the rules. Because the rules have to be changed. Everything needs to change. And it has to start today. (Greta Thunberg)

Lists

Bullet points
Precede each line in a list with a single asterisk, hyphen or plus sign.

- Land
- Air
- Water
  • Land
  • Air
  • Water

Numbered lists
Precede each line in a list with a number and full stop.

1. Land
2. Air
3. Water
  1. Land
  2. Air
  3. Water

To prevent text starting with a number being formatted as a numbered list just add a backslash.

1\. Land

1. Land

Nested lists
Indent each item in the sublist by four spaces.

1. Land:
    - African Bush Elephant
    - African Wild Dog
    - Black Mamba
    - Brown Bear
    - Cheetah
    - Coyote
    - Domestic Cat
    - Eastern Gray Squirrel
  1. Land:
    • African Bush Elephant
    • African Wild Dog
    • Black Mamba
    • Brown Bear
    • Cheetah
    • Coyote
    • Domestic Cat
    • Eastern Gray Squirrel

Tables

Divide words with hyphens to create the column names and then separate each row with a pipe. Colons are used to align the values in the cells.

|      |Column 1 |Column 2 |Column 3  | 
|:---- |:------- |:-------:| --------:|
|Row 1 |is |is |is |
|Row 2 |left |centre |right |
|Row 3 |aligned |aligned |aligned |
Column 1 Column 2 Column 3
Row 1 is is is
Row 2 left centre right
Row 3 aligned aligned aligned

R Markdown

The workflow for creating a dynamic document in R is:

  1. Create a new R Markdown document from the menu: File -> New File -> R Markdown
  2. Create a YAML header
  3. Write the content with markdown
  4. Embed the R code in code chunks or inline code
  5. Click the Knit button to compile it